Nanostructuring Effect of Plasma and Solvent Treatment on

Langmuir 2004, 20, 11183-11190

11183

Nanostructuring Effect of Plasma and Solvent Treatment on Polystyrene Elmar Bonaccurso† and Karlheinz Graf*,†,‡ Max-Planck-Institut fu¨ r Polymerforschung, Mainz, Ackermannweg 10, D-55128 Mainz, Germany, and Physikalische Chemie II, Universita¨ t Siegen, D-57068 Siegen, Germany Received December 23, 2003. In Final Form: September 9, 2004 Plasma treatment of polymer surfaces is used to control the generation of topological surface structures: stripes, starlike morphologies, and pinnacles in the range from 100 nm up to several micrometers. These protrusions arise when the plasma-treated polymer surface is exposed to an organic solvent (liquid or vapor phase). The distribution density and the height of the observed structures on the surface are functions of the power density of the plasma reactor and the exposure time to the plasma, the duration of the development process, the type of the polymer, and its manufacturing. We suggest that the structures are generated by selective swelling of less cross-linked areas within the polymer surface and not by rearrangement or dissolution of polymer chain fragments created by plasma, or by amphiphilic moieties due to oxidation as a consequence of plasma treatment.

Introduction Polymer surfaces have been used successfully in fields such as adhesion, biomaterials, protective coatings, friction and wear, composites, microelectronic devices, and thinfilm technology. In general, specific surface properties such as chemical composition, hydrophilicity, roughness, crystallinity, conductivity, lubricity, and cross-linking density are required for the success of these applications. Polymers very often do not possess the surface properties needed for these applications. However, they have excellent bulk physical and chemical properties, are inexpensive, and are easy to process. For these reasons, surface modification techniques that can transform these inexpensive materials into valuable finished products have become an important part of the plastics industry. In recent years, many advances have been made in developing surface treatments to alter the chemical and physical properties of polymer surfaces without affecting bulk properties.1 Among these, plasma treatment is probably the most versatile technique. Water and different types of gas such as argon, oxygen, nitrogen, fluorine, and carbon dioxide are used to produce desired surface properties.2-7 For example, oxygen-plasma treatment can increase the surface energy of polymers, whereas fluorineplasma treatment can decrease surface energy and improve chemical inertness. Inert gas plasma can induce cross-linking at a polymer surface as well as modification of the surface roughness. Changes by plasma treatment are usually confined to the top several hundred angstroms. * Corresponding author. E-mail: [email protected]. Phone: +49-6131-379115. Fax: +49-6131-379310. † Max-Planck-Institut fu ¨ r Polymerforschung. ‡ Universita ¨ t Siegen. (1) Chan, C. Polymer surface modification and characterization; Hanser: Mu¨nchen, 1994. (2) Clement, F.; Held, B.; Soulem, N. Eur. Phys. J.: Appl. Phys. 2002, 17, 119-130. (3) Foerch, R.; McIntyre, N. S.; Hunter, D. H. J. Polym. Sci., Part A: Polym. Chem. 1990, 28, 193-204. (4) Foerch, R.; McIntyre, N. S.; Sodhi, R. N. S.; Hunter, D. H. J. Appl. Polym. Sci. 1990, 40, 1903-1925. (5) Hopkins, J.; Wheale, S. H.; Badyal, J. P. S. J. Phys. Chem. 1996, 100, 14062-14066. (6) Briggs, D.; Rance, D. G.; Kendall, C. R.; Blythe, A. R. Polymer 1980, 21, 895-900. (7) France, R. M.; Short, R. D. Langmuir 1998, 14, 4827-4835.

Plasma treatment generally is divided into four distinct mechanisms:8 (i) ablation or etching; (ii) cross-linking or mechanical stabilization; (iii) activation/deactivation or change of surface energy; (iv) sputtering or deposition of material. Mixtures of argon and oxygen are commonly used for reactive ion etching (RIE) of the surface. During this process, the rate of surface oxidation can be controlled. This treatment has also been found to induce, or enforce, cross-linking between polymer strands.9,10 Side effects of RIE are a change of surface roughness and surface energy (usually both increase), a mechanical stabilization of the surface layers, and the formation of smaller fractions of polymer (low-molecular-weight (LMW) fragments) produced by chain scissions. The LMW fragments can easily be washed off the surface with a solvent in an ultrasonic bath.11,12 In this paper, we demonstrate that air or argon plasma combined with a solvent treatment can be used to generate stable topological surface structures in homopolymers as polystyrene (PS) and poly(methyl methacrylate) (PMMA), thereby controlling their surface roughness. This is a major issue, for example, in processes that promote adhesion between polymer and other surfaces, in bonding processes between polymers, or for the generation of superhydrophobic fluorinated surfaces.13-15 We produce the relief structures first by cross-linking the polymer in plasma and then by swelling it in an organic solvent (vapor or liquid phase). It was observed in block copolymers that swelling with selective solvents and consecutive drying leads to protrusion of the lower over the higher solubility (8) Liston, E. M.; Martinu, L.; Wertheimer, R. J. Adhes. Sci. Technol. 1993, 7, 1091-1127. (9) Grant, J. L.; Dunn, D. S.; McClure, D. J. J. Vac. Sci. Technol., A 1988, 6, 2213-2220. (10) Larsson, A.; Derand, H. J. Colloid Interface Sci. 2002, 246, 214221. (11) Murakami, T.; Kuroda, S.; Osawa, Z. J. Colloid Interface Sci. 1998, 200, 192-194. (12) Murakami, T.; Kuroda, S.; Osawa, Z. J. Colloid Interface Sci. 1998, 202, 37-44. (13) Busscher, H.; van Pelt, A. W. J.; de Boer, P.; de Jong, H. P.; Arends, J. Colloids Surf. 1984, 9, 319-331. (14) Erbil, H. Y.; Demirel, A. L.; Avci, Y.; Mert, O. Science 2003, 299, 1377-1380. (15) O ¨ ner, D.; McCarthy, T. J. Langmuir 2000, 16, 7777-7782.

10.1021/la036441o CCC: $27.50 © 2004 American Chemical Society Published on Web 11/11/2004

11184

Langmuir, Vol. 20, No. 25, 2004

phase.16,17 Here, we report a similar effect for a homopolymer. Using an extruded PS substrate with some heterogeneity in the surface properties of the polymer, a plasma treatment creates a heterogeneously cross-linked surface. A consecutive exposure to solvent causes local protrusion of the less cross-linked parts. This offers the possibility to visualize surface heterogeneities of a homopolymer with the atomic force microscopy (AFM) technique18 by topography. Alternative interpretations for the formations of such protrusions are discussed. Materials and Methods As polymer substrates, we used technical amorphous PS wafers (thickness ) 1.2 mm, F ) 1.01 g cm-3) with a broad molar mass distribution19 (Mn ) 109 000 g mol-1, Mw ) 284 000 g mol-1, Mw/Mn ) PDI (polydispersity index) ) 2.6, Goodfellow Ltd., Cambridge, U.K.). As organic solvents, we used toluene (ultrapure grade, J.T. Baker, Phillipsburg, NJ) and ethanol (ultrapure grade, Merck, Darmstadt, Germany). We cut rectangular samples (20 × 10 mm2) from the PS wafers. After the protection foil was carefully removed, the samples were directly plasma treated without any additional surface cleaning. Substrates cleaned with methanol (ultrapure grade, Aldrich) in an ultrasonic bath for 30 s before plasma treatment showed the same contact angles of sessile drops of pure water (the advancing and receding contact angles were 90 ( 2° and 78 ( 3°, respectively; DSA 10, Kru¨ss GmbH, Hamburg, Germany). For comparison between different PS samples, we also used PS powder with a molar mass in the same range as above but with a PDI close to unity (Mn ) 200 000 g mol-1, Mw ) 203 000 g mol-1, PDI ) 1.015, specification from Fluka GmbH). The PS substrates were made by adding small amounts of toluene to the PS powder in a container made of a Teflon ring with a detachable flat Teflon bottom. The toluene dissolved the PS, and by spinning at very low velocities (4-8 rpm) we obtained circular and flat PS films (thickness ) 100-200 µm, diameter ) 15 mm, root-meansquare (rms) surface roughness ) 1.6-2.5 nm). To check the influence of the preparation method for the technical PS wafer (Goodfellow Ltd.) upon the structure formation, a part of it was spin-cast from toluene solution, as we did with the PS powders, obtaining a circular and flat PS film (thickness ) 150-300 µm, diameter ) 15 mm, rms surface roughness ∼ 3.1 nm). We extended our investigations to other technical products such as 0.6 mm thick, amorphous PS substrates cut from the bottom of Petri dishes (Greiner GmbH, Frickenhausen, Germany) and 1 mm thick amorphous PMMA wafers (F ) 1.19 g cm-3, Goodfellow Ltd., U.K.). All samples were treated in an air or an argon (purity, 4.8) atmosphere in a plasma reactor (40 kHz, PlasmaPrep2, GaLa Instrumente, Germany). Two parameters were varied during the experiments: the etching power and the etching time. In a set of experiments, we varied the etching time between 0 and 600 s in 120 s time intervals, fixing the power at 30 W. We generated five ∼4 mm wide stripes on the sample surface by covering the area with a glass slide. Each stripe was etched for a different time. Between subsequent etching steps, the sample was taken out of the reactor and the glass slide was moved to a new position in order to confirm that the intermediate exposure of the plasma-treated surface to the air had no effect on the final structure. In another set of experiments, the power was varied between 0 and 60 W in 20 W intervals, fixing the process time to 120 s. Again, glass slides served as protection during the plasma treatment. In both arrangements, the local power density was difficult to estimate due to the geometry of the reactor chamber and of the electrodes. However, the samples were always (16) Elbs, H.; Fukunaga, K.; Stadler, R.; Sauer, G.; Magerle, R.; Krausch, G. Macromolecules 1999, 32, 1204-1211. (17) Elbs, H.; Drummer, C.; Abetz, V.; Krausch, G. Macromolecules 2002, 35, 5570-5577. (18) Binnig, G.; Quate, C. F.; Gerber, C. Phys. Rev. Lett. 1986, 56, 930-933. (19) Measured by Sandra Seywald and Beate Mu¨ller, Max Planck Institute for Polymer Research, Mainz.

Bonaccurso and Graf positioned in the same place, allowing thus for reproducible results under similar experimental conditions. After the plasma treatment, we “developed” the sample surface according to either one of two procedures: (i) by exposing the sample to a saturated toluene vapor in a sealed vessel; (ii) by immersing the sample into an optimized solution of 25/75 (v/v) of toluene/ethanol, usually for a time between 60 and 150 s. After the development process, the solvent had to be extracted from the polymer. We applied either one of two “drying” procedures: (i) a soft nitrogen flow, followed by exposure of the substrates to an atmosphere of nitrogen or dry air for 30-45 min; (ii) a methanol bath for 30 min to allow the toluene to diffuse out of the polymer matrix as suggested by Okubo et al.20 This is followed by drying with a soft nitrogen flow for 60 s. All surfaces were characterized with an atomic force microscope in contact and friction mode in air (Bioscope, Veeco Instruments, Santa Barbara, CA). If not otherwise stated, scan sizes varied between 25 × 25 µm2 and 75 × 75 µm2. Scan force was minimized, and the scan speed never exceeded 2.5 lines per second. AFM images show the height information data, where the darker color maps lower altitude and brighter color maps higher altitude. The images were not processed further in most cases, except for a planefit of first order, which was applied in single cases to correct for the tilting of the sample. Friction mode images were acquired both in trace and retrace mode. These images only give qualitative information and permit one to distinguish between “higher” and “lower” friction areas on the surface, because no torsional spring constant calibration of the cantilever was performed.

Results and Discussion Structure Formation on PS Substrates. For a PS substrate cut from the wafer and treated with air plasma for 100 s at 30 W, we found a pronounced formation of stripes and starlike structures on the area that was not protected by a mask (Figure 1a). The stripes are of different lengths, with sharp kinks, but partially aligned parallel. The distance between the stripes lies in the range of 0.61.1 µm for all PS samples under investigation. The starlike structures (“stars”) represent a junction of several stripes. They are usually higher than the stripes converging to or originating from them. Their height goes from several hundreds of nanometers up to 2 µm. To analyze if the plasma treatment (denoted as RIE in the following) or the following solvent treatment (“development”) was responsible for the structure formation, we acquired AFM micrographs of the PS surface before and after plasma treatment. The PS surface appeared unstructured, but a change of surface roughness could be detected in those parts exposed to plasma (1.6, 14.1, 18.0, and 21.1 nm rms roughness for samples treated with a power of 0, 20, 40, and 60 W for 120 s, respectively). Moreover, in AFM friction mode in air a distinct contrast between areas in these parts and those covered previously by the mask became evident. This contrast occurs due to different hydrophilicity, as has been observed by DupontGillain et al.21,22 It can be explained by a different friction between the AFM tip and the sample surface, depending on the degree of oxidation. This will be explained in more detail later. On the other hand, developing a fresh PS sample without plasma treatment in toluene vapor or a mixture of toluene and ethanol 25/75 (v/v) left behind a homogeneous surface. Previous surface features were smoothed out by solvent (20) Okubo, M.; Konishi, Y.; Takebe, M.; Minami, H. Colloid Polym. Sci. 2000, 278, 919-926. (21) Dupont-Gillain, C. C.; Nysten, B.; Hlady, V.; Rouxhet, P. G. J. Colloid Interface Sci. 1999, 220, 163-169. (22) Dupont-Gillain, C. C.; Adriaensen, Y.; Derclaye, S.; Rouxhet, P. G. Langmuir 2000, 16, 8194-8200.

Nanostructuring Effect of Plasma on Polystyrene

Langmuir, Vol. 20, No. 25, 2004 11185

Figure 2. Smearing out of the original surface defects due to exposure to solvent. (a) AFM scan of a PS plate as received from the manufacturer, with “grooves” and “scratches”; (b) AFM scan of a PS sample which has been exposed to a 25/75 (v/v) solution of toluene/ethanol for 90 s. Note: the vertical scale for all pictures is 50 nm.

Figure 1. (a) AFM image (contact mode, 60 × 60 µm2) on a PS substrate after plasma treatment in air (30 W, ∼100 s), development in 25/75 (v/v) of toluene/ethanol for 120 s, and drying under a nitrogen flow. On the left part of the plate, a gold mask was evaporated before plasma treatment. (b) The height profile is taken along the dashed line. Bright parts in the AFM image are higher than dark ones. The arrows indicate the highest peaks on the surface.

treatment. Such features are for example characteristic lines on the original PS surface (Figure 2a) originating from the manufacturing process of the plate (extrusion23). The rms roughness was 3 ( 1 nm for the fresh PS surface and 2 ( 1 nm for the sample after exposure to solvent (Figure 2b). These observations lead us to the conclusion that the topological pattern generation in extruded homopolymers only occurs due to the combination of RIE and exposure to solvent, and this, to our knowledge, has never been described. Before we go into details of the structure formation, we will first present our study on the dependence of the (23) Personal communication by Paul Everitt and Barry Saunders, Goodfellow Ltd., U.K.

Figure 3. Dependence of PS surface patterning on argon plasma treatment time (power ) 30 W). Samples were developed in a 25/75 (v/v) solution of toluene/ethanol for 90 s, the solvent was extracted in methanol for 30 min, and some samples were dried under a soft nitrogen flow. The vertical scale for all pictures is 1500 nm. The boxes are the areas from which the roughness for Figure 5a is taken. Note: Dark is high, and bright is low.

formation of these structures on the exposure time to plasma and the power of the applied plasma. Structure Formation: Time and Power Dependence. A pronounced change in the appearance of the protrusions with exposure time and power in an air or argon plasma could be observed (Figures 3 and 4). For the time dependence (Figure 3), the AFM images were inverted to better visualize the action of the plasma; the bright parts are lower than the dark ones. The area of the lower parts increases from about 50% after 120 s to about 95% after 600 s. The number of the dark appearing “stars” decreases in the same time interval, while the “stripes”

11186

Langmuir, Vol. 20, No. 25, 2004

Bonaccurso and Graf

Figure 5. (a, top) Dependence of the rms roughness on the argon plasma treatment time (power fixed to 30 W) and power (time fixed to 120 s) as taken from Figures 3 and 4. The dashed line represents the roughness (black squares) of the structureless parts in the boxes of Figure 3c-e. (b, bottom) Qualitative cartoon (cross section) of protrusions developing within the plasma-treated PS surface after exposure to toluene (increasing plasma power from left to right). For details, see the text. Figure 4. Dependence of PS surface patterning on argon plasma power. The etching time for all samples was 120 s. Samples were developed in a 25/75 (v/v) solution of toluene/ ethanol for 90 s, the solvent was extracted in methanol for 30 min, and some samples were dried under a soft nitrogen flow. The vertical scale for all pictures is 500 nm, except for 0 and 60 W, where it is 50 nm. Note: Dark is low, and bright is high.

disappear already after 360 s. From this time on, large areas of the substrate appear flat, whereas even after 10 min of plasma exposure some of the starlike protrusions still persist. With the disappearance of the stripes between the stars after 360 s, the original lines on the PS substrates (see also Figure 2a) remain visible. We obtained similar topological patterns varying the etching power between 0 and 60 W (Figure 4). Plunging the PS into the toluene/ethanol solution without previous argon plasma treatment resulted in a uniform surface. A moderate power of 10 W resulted in the formation of protrusions on the surface with small diameters (1-4 µm). At a power of 30 W, patterns of the original surface became visible as lines amidst the protrusions (three of these lines are highlighted in the picture). These were larger than in the previous case (4-8 µm). The general tendency of surface expansion perpendicular to the PS surface is nearly suppressed at 40 W. Additionally, the protrusions are much shallower and the original line structure becomes more visible than at 30 W. Singular stars appear randomly. Their height from the bottom varies between 300 and 1500 nm. Treating the PS with plasma for 120 s at 60 W resulted in a conservation of the original line structure without star- or stripelike protrusions. We also compared the height evolution within the sample after exposure to plasma and solvent with an area that was solely exposed to solvent. Generally, we found a decreasing tendency of

height changes in the surface with increasing exposure time to plasma or plasma power. From the original images, we determined the variation of the surface roughness (rms roughness, Figure 5a) according to plasma-treatment duration and power. The plots show a maximum at 240 s and 30 W, respectively, indicating the most pronounced structure formation in the polymer surface. The roughness increases from a few nanometers to 150 nm. After long exposure times and for high power, the roughness decreases again to approximately 10 nm (dashed line). This value can also be found after shorter exposure times to plasma in those parts which lie between the starlike protrusions (Figure 3, boxes). This shows the equivalence of plasma power and its exposure time on the modification of the polymer surface. Summarizing these findings, we draw a cartoon, which illustrates the structure formation in the extruded PS surface (Figure 5b) based on the AFM images from Figure 4. Since we applied different power levels on the same PS substrate by the technique described in the experimental section, we can directly compare the height of the protrusions for different power levels. Especially, it is possible to compare the parts that were subject to maximum expansion after exposure to solvent (100%, leftmost) and those that were conserved completely without any tendency to expand upon exposure to solvent (0%, rightmost). This is the level of the original PS surface. The cartoon can be thought of as a cross section over the whole length of the PS substrate after plasma treatment and exposure to solvent. We can distinguish five phases in the evolution of protrusions:

Nanostructuring Effect of Plasma on Polystyrene

(i) Without plasma treatment, the polymer surface expands the most and completely loses its original surface structure (e.g., lines). (ii) Short exposure times or low power (120 s, 10 W) generates small protrusions. The average change in height is ∼90% of the maximum height expansions for a fresh sample, but some pinnacles may reach 100%. The original surface structure is lost. (iii) Medium exposure times or power (240 s, 30 W) generates large protrusions. The average change in height is 30-50% of the maximum expansion, but some peaks may exceed it. (iv) Long exposure times or high power (360 s, 40 W) generates a relatively uniform surface with single, big protrusions decorating the surface. The average change in height is only 5-10% of the maximum expansion. The original surface structure of the fresh sample becomes visible. (v) Very long exposure times or very high power (600 s, 60 W) leads to a largely uniform surface without protrusions. The original surface structure is preserved. Structurally, we can distinguish three main features. Lines from the fabrication process appear after roughly 360 s of plasma treatment or at a power higher than 30 W. Stars occur between 0 and 120 s or at a power larger than 10 W but vanish at very high power or after long exposure times. Stripes occur for low power or short times in plasma but largely vanish already after roughly 360 s. They persist longer in the vicinity of the stars. In the next section, we will discuss the structure formation in detail. Discussion of Structure Formation. First of all, in the original extruded PS surface we observed lines, which vanish upon exposure to solvent. Solvent uptake leads to swelling. It is reasonable to assume a surface rearrangement upon swelling such that the original fabrication lines disappear. The increasing conservation of these lines with time and increasing power clearly shows that the argon plasma stabilizes the PS surface, probably by cross-linking the polymer surface. From the height of the fabrication lines (